The PMAR project - European Commission · The PMAR project: Piracy, Maritime Awareness and Risks...

13 June 2012 GeoMaritime, London 1

The PMAR project

European Commission – Joint Research Centre

Mombasa, 21 Nov 2012 PMAR Workshop 2

Illegal immigrationIllegal immigration

Oil pollutionOil pollutionPiracyPiracy

Maritime TrafficMaritime Traffic

Maritime SecurityMaritime Security

Maritime SafetyMaritime Safety

Illegal fishingIllegal fishing

Maritime Awareness

To govern our seas we need: Maritime Situational Awareness

• Which ships are on the sea?• What are they doing?• Do they pose a threat or

problem?


The PMAR project: Piracy, Maritime Awareness and Risks

• Build up Maritime Awareness

• As a key enabler for fighting piracy• Practical, cost-effective

• In the hands of maritime authorities in the region• Tailored to the region’s needs

• In a regional approach• In a civilian, data sharing context

• Beyond piracy, also for wider maritime safety and security issues• Illegal fishing, immigration, pollution, etc.

• Toward better maritime governance in general


Contents

1. Introduction – What is PMAR2. Overview of maritime surveillance tools3. Details on surveillance tools4. Data processing, fusion5. Trials6. Performance7. Trials – observation data8. Conclusions and way ahead


PMAR

The PMAR project

• Requested by European Parliament• Executed by European Commission

• Overseen by EuropeAid• Technical work by JRC

• 2010-2012• Part of the EU’s Regional Maritime Capacity Building initiatives in the Horn of Africa

• Based on needs assessment from visits

1. Introduction – What is PMAR


Embedding of PMAR

• Part of EU initiatives in HoA• Instrument for Stability / Critical Maritime Routes• MARSIC• MASE• EUNAVFOR Operation Atalanta• EU Training Mission (EUTM) in Somalia• EUCAP Nestor• Others: PAPP, GMES Africa, …

• Closely coordinated with • IMO’s counter-piracy activities• Counter-piracy activities from other countries• At global level, with CGPCS


Tools for maritime surveillance

1. From the coast•Looking out•Cameras•Radar

Libreville, Gabon, Feb 2012

2. Overview of surveillance tools



1. From the coast•Looking out •Cameras•Radar•AIS messages from shipsOut to some 30 km With AIS, ship

continuously broadcast their position on VHF, for all to receive. Globally mandated by IMO on big (>300 tons) ships.

Coastal AIS is inexpensive and easy to install and operate. The MSSIS / SeaVision network enables sharing of the picture



1. From the coast•Looking out •Cameras•Radar•AIS messages from shipsOut to some 30 km

2. Going to sea•Patrol boat•Maritime Patrol AircraftExpensive, but also allows intervention

Mauritius Coast Guard


3. Ocean-wide•LRIT

• Merchant ships (>300 ton) report every 6 hours to their Flag State, globally mandated by IMO

•VMS• Medium and big fishing ships report every 2 hours

to their Flag State, mandated at national level•AIS received by satellite

• From coastal to global coverage•Earth observation from satellite

• To also find non-reporting ships


www.worldfishing.net

LRIT


Satellite radar image (SAR)

RADARSAT-2 satellite© MDA 2011


Maritime monitoring & surveillance tools - Summary

• Ship reporting (cooperative) systems• LRIT, AIS, VMS, call-in/radio

• Observation (non-cooperative) systems• Camera, infrared camera, radar, sonar

• Platforms for observation systems• Shore, sea bottom, ship, airplane, UAV, satellite

• Interception• Patrol boat

• Auxiliary tools• Maps, ship registers, weather information


PMAR study questions

• Which of these systems to use?• Suitability for counter-piracy• For use in the region, by authorities in the region• For civilian / constabulary tasks• Operational readiness• Cost - benefit

• What products are required? • How can we make that product? • How good does it finally satisfy the needs?


Products

• Maritime Situational Picture• Map showing the ships where they are currently

• Historic maps of ship traffic• Describing what is normal traffic• To enable finding abnormal behaviour

• Historic maps of piracy incidents• To localise risks

• Predictions• Ship positions, weather

• Based on the above: Actions can be taken• Issue warnings, general or specific• Plan routes• Plan patrol assets, do interceptions


Black: Coastal AIS MSSISBlue: Sat-AIS LuxSpaceGreen: Sat-AIS exactEarthRed: LRIT (IT)Gray: 3-8 hr awayLight gray: 8-24 hr away

movie

Maritime Situational Picture – Reconstructed


PMAR system

PMAR system Maritime awarenessActionable informationInput data

3. Details on surveillance tools


LRIT

• Long Range Identification and Tracking• Designed for maritime security

• Globally mandated by IMO in SOLAS convention• Cargo ships > 300 GT, Passenger ships > 12 pax,

Mobile off-shore drilling units; on int’l voyage (…) • Message with ship ID and position sent 6-hourly

to Flag State (Maritime Authority)• Frequency may be increased up to every 15 min• Position request possible

• Global coverage through satellite communications• The reports may go to:

• Coastal State within 1,000 NM• Destination Port State• But for a fixed per-message cost


AIS

• Automatic Identification System• Designed for maritime safety (collision avoidance)

• Globally mandated by IMO in SOLAS convention• Passenger ships, Ships > 300 GT on int’l voyage,

Cargo ships > 500 GT, (…) • Message with ship ID, position, speed, course,

ship & cargo type, navigation status, port & time of destination, …, broadcasted every few sec - min• Frequency depends on speed & manoeuvre

• Broadcasted on VHF (to the horizon or bit beyond)• The reports are received by:

• Other ships in the vicinity• Coastal stations


Satellite AIS

• Reception of the AIS messages from space• Provides global coverage

• However doesn’t work well in dense shipping• Comes with some delays (downlinking to Earth)• Not continuously updated but only when satellite

passes over• So mostly suitable for

remote areas• Can be done by small,

inexpensive satellites• Has quickly become

commercially available


VMS

• Vessel Monitoring System• Designed for fisheries control

• Nationally regulated• On the bigger fishing ships• Global guidance by UN-FAO

• Message with ship ID, position, speed & course sent 2-hourly to Flag State (Fisheries Authorities)• Frequency may be increased• Position request possible

• Global coverage through satellite communications• The reports may go to:

• Coastal State within its EEZ• Secretariats of Regional Fisheries Management

Organisations


Optical•Only daytime, clear skies•Better for details

Use for recognition

Radar•Independent clouds, night•Better for wide areas

Use for detection

Cargo ships in IstanbulIKONOS satellite

Line of fishing vessels in NE AtlanticRADARSAT satellite

Satellite images: Optical or Radar


• Suitable for use on satellite (long range)• Wide area, low resolution Small area, high

resolution

Synthetic Aperture Radar (SAR)

Mombasa, 21 Nov 2012 PMAR Workshop 23ENVISAT ASAR AP

Synthetic Aperture Radar images

Radarsat‐2 UltraFine

HH

TerraSAR‐X Stripmap

Medium resolution (25 m)

High resolution (1‐3 m)

Satellite SAR


Resolution15-25 m

Swath100-200 km

Medium resolution wide swathShips

BigMedium

Small

20 m fishing boat


Radarsat-1 ScanSAR Narrow A 20 hours later

Envisat Wide Swath VV

Satellite SAR also suitable for oil spill detection


SPOT-5 10 m color

SPOT footprint, 60 x 60 km

Optical satellite images


SPOT-5 2.5 m b/w

EROS1.8 m b/w



Fishing vessel

IKONOS

Ships at oil well



Pleiades-1 new optical satellite

10 x 20 km swath50 cm resolution

…but there are also optical satellites with

600 km swath and 22 m resolution (DEIMOS‐1)

launched Dec 2011


Space-based surveillance limitations

• No continuous monitoring possible, only periodic updates• Sat-AIS: roughly every 6 hr now, will go down to

sub-hourly• Sat images: roughly daily,

but have to select area• Data are available with a delay

• Need to be downloaded to Ground Station, processed, sent to user

• 2 hour – 15 min• Data have a cost

• Considering the information content, satellite images are more expensive than satellite AIS


Space-based surveillance limitations (II)Sat imaging only:• Need to plan data request days in advance

• Select from available satellite overpass times• Choose between

• Wide area (300 km) – big ships (> 35 m)…

• Small are (10 km) – small boats (> 5 m)• Ship identification not possible

• Detection, classification• Classification much easier with optical images (but

clouds-limited)

Sat-AIS only:• Don’t find not-reporting vessels


Auxiliary data

• Maps• Coastline maps• Digital Nautical Charts

• Ship registers• Link ship IDs

• IMO number, MMSI number, call sign, name

• History, ownership, …• Weather data

• Meteo• Wind, clouds, visibility

• Oceanographic• Waves (height, length, direction, steepness),

Currents, Fronts


Choosing data sources (based on practical, affordable, regional, …)

• Reporting systems• Coastal AIS, government operated• Satellite AIS, from several commercial providers• LRIT from own Flag, and others up to 1,000 NM• VMS from own Flag, and others in own EEZ• Call-in reports from ships (specific areas / lines)

• Observation systems• Satellite SAR• Satellite Optical (special locations)

• Auxiliary data• Maps, Ship registers, Met/ocean data

• Not in PMAR (regional concept) but also needed• Coastal radar (+ cameras & sightings)• Patrol assets (boats, aircraft)


PMAR system

PMAR system Maritime awarenessActionable informationInput data

4. Data processing, fusion


SHIPREGISTER

MARITIMESITUATIONAL

PICTURE

VESSEL TRACKING

SHIPTRACKS

FUSION ALGORITHM

identifier name flag Type length247001002 PIPPO ITA CARGO 250

COOPERATIVESHIP POSITIONS(AIS, LRIT, VMS)

NONCOOPERATIVE

SHIP POSITIONS(VDS)

Building the Maritime Situational Picture


Tracking

• The track is the series of positions of one ship• Use the MMSI number as primary ship ID

• 9-digit number, first 3 is country (e.g. 241001002)

• Issues: • In LRIT, only IMO number is given

• Convert IMO MMSI with Ship register• Sometimes the relationship changes

• Errors in messages (at transmission or reception)• Remove outliers, recognise strange values

• Multiple use of one MMSI number• Track separation


Vessels sharing MMSI number


GatingFor land crossing

x

Multiple-used MMSI tracking


GatingFor max speed



Gating

Correlation

x



Gating

Correlation

Track initiationConfirmationDeletion

Prediction



Prediction

• Maritime Situational Picture is at certain reference time (now)

• Predict all ship positions to MSP time, based on their positions known so far

• Use speed, course in message; or most recent two positions

• Issues: • Maximum prediction time, then ship ‘lost’• Ship may change course

• The longer ago the most recent messages, the less reliable the prediction!

• Errors in speed, course• Errors in time (not in AIS message, externally

affixed)


Communication

• All information is generated at different locations• It has to come together in one place• Results have to go to the end users

• Upstream (from the individual sensors)• Radio links, satellite communications• Data collected by satellite come with a delay

• Downstream (to us)• Web-based, for the most part

• TCP/IP, FTP, HTTPS

• Radio / satcom for call-in reports


PMAR trials

• Two trials were run• To develop and to verify the concept and the system• Collection of reporting data over wide area during

several months period• Collection of satellite images focussed on some

areas, mostly the Gulf of Aden• End 2010

• To design the system• Aug 2011 – Jan 2012

• To verify and fine-tune the design, evaluate performance, obtain cost-benefit data

5. Trials


Choosing data sources• Reporting systems

• Coastal AIS, gov’t operated• Satellite AIS, from several commercial providers• LRIT from own Flag, and others up to 1,000 NM• VMS from own Flag, and others in own EEZ• Call-in reports from ships (specific areas / lines)

• Observation systems• Satellite SAR• Satellite Optical (special locations)

• Auxiliary data• Maps, Ship registers, Met/ocean data

• Not in PMAR (regional concept) but also needed• Coastal radar (+ cameras & sightings)• Patrol assets (boats, aircraft)

From EU Flags

Used in trial


PMAR data sources used in trial

• Reporting data• LRIT – EU Flags• AIS

• Coastal AIS – MSSIS• Satellite AIS – exactEarth, LuxSpace/OrbComm,

FFI (NORAIS, AISSat-1)• VMS (fishing ships) – few EU Flags– Continuous data stream, update minutes – hours• Incident reports – IMB, ReMISC Sana’a

• Satellite bservations (non-cooperative)• SAR radar (medium & large ships) – ENVISAT (ESA),

RADARSAT-2, TerraSAR-X, CosmoSkymed• Optical – SPOT, IKONOS, DEIMOS (EUSC)– Infrequent snapshot images

• Auxiliary• Ship registers• Coastline maps

CommercialInstitutional

Cooperations with: EMSA, EUSC, NURC, Italian Coast Guard, EU Flag States, FFI (NO), ESA, SPAWAR, US Navy, DLR (GE), GD (US), …


Trial results – Ship reporting data

Last positionson 2 Dec 2011

green = FFIyellow = exactEarth

cyan = LuxSpacered = MMSIS


Example

Liquid Velvet, MMSI 538003117, chemical tanker, reporting navstat 1 “at anchor” outside Somalia. Pirated 2 November 2011


Example

Support and combat vessel Patino, MMSI 224503000, part of a EU Atalanta. On 8 Dec 2011 observed 5 times, steaming at 18-20 knots, toward pirated ships seen in the lower part. This is the ship that was attacked by pirates on 12 Jan 2012 http://www.eubusiness.com/news-eu/spain-africa-piracy.eij/

http://www.eubusiness.com/news-eu/spain-africa-piracy.eij/


Example

GOLDEN LAKE NO.808, MMSI 440916000, observed 15 times

Two Korean trawlers operating for the whole day along a 9 nautical miles long part off the Somalia coast.


A ship track between 1 and 14 Dec 2010Positions from LRIT (red), MMSIS (black, near Mombasa, Yemen), LuxSpace (green) and exactEarth (blue).

The lack of AIS points can be explained by the standing recommendation to switch off AIS in pirate-risk areas.

Ship traffic fused from different sources – examples


Ship traffic fused from different sources – examples

One ship. Latitude vs. TimeColour per Sat-AIS platform


Distribution of speeds


Sat-AISExactEarth

tracks per day

993 different ships

Data source: exactEarth

11 Dec 2010

Ship tracks from Satellite AIS


Performance / added value

• Performance measures• How many different ships are seen?• Geographical coverage• Update rate• Accuracy of positions• Accuracy of other info

• Some info in AIS messages is entered by hand

• Added value• If we use N data sources, how much more

information do we get from N+1 sources?

• This is evaluated using the 2011 trial data in the period 15 Nov – 15 Dec

6. Performance


AIS messages per day

On average 48,600 messages/dayFluctuation due to availability of satellites and connections


Satellite AIS coverage

1 month7 systems

1 day7 systems

1 day1 satellite

1.5 million ship position reports

Avg 1,200 ships/dayAvg 40 reports/ship/day


AIS coverage – 1 month

MSSIS

1 satellite

All systems


# Messages per vessel

All systems1 month


Distribution of update times per vessel

single satellite


Distribution of update times per vessel

MSSIS

single satellite

All systems


# Messages / ship on one day



16 updates, 218 messagesSEVEN SEAS HIGHWAY, MMSI 353098000,18 kn



2 updates, 3 messagesNORNA N, MMSI 353217000, tanker


Added value: # Distinct ships

Provider 1 2 4 5 1&2 1&2&4 1&2&4&5Average 339 620 901 1008 783 1053 1183Increase 444 270 130Stddev 84 516 57 68 370 144 74

1 provider=

2 satellites


Added value of extra providers

New obs=

after> 30 min


Accuracy of the MSP

• The MSP shows all ships at the same time• Positions are predicted from last reported ones• May be minutes or hours old• How can we assess the accuracy?


Accuracy of the MSP

• Predict to each reported position, using previous positions• Compare with reported position


Ship position accuracy in MSP

Systems # MMSI in whole

month

# Ships at one time

(mean)

Mean position

error (NM)

% Position errors

< 10 NM

MSSIS 1,851 110 8.4 82

Single Sat-AIS (mean)

4,363 607 10.5 74

Four Sat-AIS + LRIT + MSSIS (mean)

5,022 931 5.9 86

All 5,155 1,011 5.0 88

Allno prediction

5,155 1,011 23.3 58

+ 1.4 NM / hour delay


Finding the non-reporting ships

Non-reporting ships: •Most ships < 300 GT

• Some of these carry AIS voluntarily• In particular, Class-B AIS

•LRIT ships under a non-cooperating Flag• Within 1,000 NM, (almost) all can be bought• PMAR trial has 34 Flags• IMO Piracy polygon has 87 Flags

•Ships with AIS that have it turned off• BMP in high piracy risk areas

How to find them?•Coastal radar – only out to about 20 nm•Airborne patrol – detailed info •Satellite imaging – wide area

7. Trials – Observation data


Ship surveillance from imaging satellites

next slide…

Mombasa, 21 Nov 2012 PMAR Workshop 742010-12-05 02:52

TerraSAR-X

Sat-SAR + Sat-AIS data fusion

NORAIS

2010-12-05 08:27

exatEarth

2010-12-05 07:55

AISSat-1

2010-12-05 06:50

NORAIS

2010-12-04 22:41

2010-12-04 21:23

exactEarth

2010-12-04 21:08

NORAIS

Predicted convoy configuration


Ships identified in Satellite SAR image using Satellite AIS


2011 campaign

reporting 460not reporting 571

Satellite SAR images and ships detected (Nov-Dec 2011)


2011 campaign

• green: reporting• red: not reporting (55 %)

Ships detected in satellite imagesvs. reporting ships (Dec 2011)

10 days in the Gulf of Aden


Cutlass Express, 25-27 Oct 2011

Ships (pink) detected from satellite image (TerraSAR-X, overpass 25 Oct 03:05 UTC)

seen in SeaVision at 17:18 UTC that day

Satellite images 25-27 Oct


Ship density map


Source: IMB

Historic piracy incidence map


Use for risk assessment

Ship track

Piracy incidents 2011


Technical conclusions

• Continuous (partial) surveillance of coastal zone is easily possibly with AIS• Low-cost, easy to maintain coastal AIS receivers

• Recent tools (LRIT, Satellite-AIS) enable Maritime Situational Awareness over wide areas continuous, but partial (only reporting ships)

• The continuous picture from the reporting ships can be complemented with detections from satellite imaging, on an incidental basis

• These technologies are accessible to operational users in the East Africa / Indian Ocean region• For counter-piracy, but also illegal fishing, immigration, …• Further capacity building efforts are needed

• Finding (+ tracking, + classifying) small boats remains a challenge• Satellite images can help a bit in small, targeted areas

8. Conclusions and way ahead


Possible way ahead

• Set up a PMAR-like functionality in the region

What would be needed?• An operational centre in the Horn of Africa /

Western Indian Ocean region• Staff who use it

• Implementation of the system there• Including capacity building

• Purchase of commercial data (sat-AIS)• e.g. on a yearly contract

• Building up trust with the international community• Use of data that is normally not shared


Possible way ahead – implementation

Implementation aspects• Gradual and evolutionary• Phased transfer of functionality from EU to Africa• Co-development, based on user feedback• Parallel with capacity building

• Training for use• Setting up technical support community

• IT maintenance, R&D

• The complete functionality could be commercially bought, as a service or as a system

• But that would only be meaningful if capacity building is fully included in the purchase


Options in implementation

• Geographical coverage• Western Indian Ocean Basin, or• Critical High Seas routes, or• EEZ (200 NM), Shelf (350 NM)

• Number of satellite AIS suppliers• Trial: Two suppliers (4 satellites) is not bad• NB: Suppliers promise to increase # satellites

• LRIT • At standard LRIT cost, or• International agreements for use in a regional centre

• Number of centres• One for whole region• A few for sub-regions


Options for implementation – mid-term

• Take in call-in reports from ships• Needed: Acceptance by shipping community

• Integrate the use of meteo and ocean data• Better risk assessment

• Two level system• One common regional picture• Enriched in each zone/country with local data

• These could be points of co-development during implementation


Options for implementation – Further away

• Structurally include VMS• Better protection of fishing ships (against pirates)• Better protection of fish stocks (against fishing ships)!• Probably needs involvement / setting up of

Regional Fisheries Management Organisation• Existing: IOC, IOTC, SIOFA

• Couple with national coastal surveillance systems• Not all local ship movements, but only:

• Suspicious ships• Ships moving from one zone to another• Ships on request

• This kind of data sharing is now being discussed within the EU!


www.jrc.ec.europa.eu

Contact: [email protected]

Serving societyStimulating innovationSupporting legislation

Joint Research Centre (JRC)

THANK YOU

The PMAR project - European Commission · The PMAR project: Piracy, Maritime Awareness and Risks...

Documents

Transcript of The PMAR project - European Commission · The PMAR project: Piracy, Maritime Awareness and Risks...